Chapter 19. The read

A working draft, November 2026. This book is being written while The Grove is being built; expect it to change. Checked against the build as it stood on 1 November 2026.

In this chapter

• Why dashboards and watching people leave an organisation blind to itself
• What the read looks at: a ledger of work at each branch, with no place for a name
• How work is weighed by whether the world asked, and how grow, wither, split and merge are read
• How tensions land on the right forum, and how the lean shows drift and the unspoken
• What the flame says about each branch, and why no reading can decide, pay or seat anyone

Here's the doubt, put the way a tired practitioner puts it.

You can't see what's really going on in an organisation. Not from above, and not from a screen. Dashboards measure what's easy to count, so they show tickets closed and hours logged and calls answered, and miss the conversation that saved a client. And anything built to watch the work ends up watching people, because people do the work. Give it a year and the tool that was meant to help you understand the organisation is the tool your colleagues learn to perform for.

That's not cynicism. It's experience. Most of us have watched a well-meant measure turn into a game.

What watching costs

The economist Charles Goodhart noticed it in monetary policy in the 1970s, and the anthropologist Marilyn Strathern put it in the form most people now quote: when a measure becomes a target, it ceases to be a good measure.¹ A team told to close tickets closes them, whether or not the problem is solved. A department told to hit its objective by the quarter hits it, and quietly lets go of whatever the objective didn't name. Objectives and key results, meant to align effort, often become one more number to be seen to move.

The cost to the business is that the organisation goes blind while looking at itself. The numbers say things are fine right up to the week they aren't. The cost to people is worse. Tools that track activity per person, keystrokes, response times, who was online, teach everyone that the safest thing is to look busy. What they notice and don't say, the part of the work that never fits a field, goes underground. Honesty is the first casualty of being watched.

So most self-managing organisations do the sensible thing. They refuse the dashboard and rely on conversation: check-ins, retrospectives, a sense of the room. That's better. But a room only knows what the people in it are willing to say, and it forgets. The tension everyone felt in March is gone by June.

Looking at the work instead

The psychologist Karl Weick spent much of his career on how organisations make sense of what's happening to them. One of his insights was that sense isn't found lying about. People make it, together, from cues they notice, and they make it mostly after the fact.² The cues matter. If the only cues an organisation has are what people chose to report, its sense will be as partial as the reporting.

Stafford Beer, designing his model of a viable system in the early 1970s, made a related point from cybernetics. A viable organisation needs a channel that looks at operations directly, not through the reports of the people being looked at.³

Put those together and a different question appears. Not how is everyone doing? but what is the work doing? Work leaves traces whether anyone reports it or not. It arrives, waits, gets passed on, gets done, gets turned away. Those traces can be read without asking anyone to perform, and without anyone being the subject of the reading.

What's read, exactly

In an organisation on The Grove, every piece of work that moves leaves an entry in a ledger of flows. Each entry has three things: the day, the branch, and what happened. Arrived. Done. Turned away as not ours. Passed to another branch. Refused. Waiting with no home. Done from outside the seat it belongs to, in which case the entry records the branch whose seat the doer holds. That's the whole of it. There is no field for a person. It isn't that names are stripped out afterwards; the ledger has no place to put one.

Not every piece of work counts. Each is weighed by how it crossed the organisation's membrane, its boundary with the world. Work that came from outside is the weightiest. So is the organisation's held work: what its own jobs, measures, decisions and reviews ask of it. Work going out, a post or a message sent first, buys no weight: it asks, and nobody has answered yet. Work made here, added by hand with nothing outside behind it, buys none either. Only work from outside and held work can make a branch want to grow, and work from outside only if it cost the sender something real. A like or a passing comment shows attention, not need. An enquiry, a booking, a payment, a conversation: those count. Demand for something beside the organisation's purpose grows nothing on its own; whether to grow towards it is a question for the forum at the root.

So nobody can grow their part of the organisation by keeping busy. The world has to ask. And a branch where most of what's on it was made inside gets a gentle question of its own: most of this was made here, with little from outside behind it; what is it for?

The read looks back four weeks. Recent work weighs more: something that happened today counts in full, something a fortnight old counts half, something four weeks old a quarter, and anything older drops out. Nothing is stored as a verdict. Each reading is worked out fresh when someone looks, and a tension fades when the work stops showing it, without anyone having to close it.

Grow, wither, split, merge

From those traces come four readings about a branch's shape.

A branch wants to grow when work arrives faster than it closes: enough recent arrivals to weigh about two and a half, and more than half again what was closed. Work turned away as not ours counts as closed, so a branch saying no to what isn't its job never reads as falling behind. Growth also shows when the branch's work keeps being picked up from someone else's seat, or when work with no home keeps waiting there.

A branch may be withering only when two things are true together. It's quiet: nothing has arrived or been done for four weeks, and it's past its own first four weeks, so a new branch can't wither before it starts. And something has been left untended: none of its accountabilities touched, its measures unread for thirty-five days, or its seat empty for fourteen. Quiet alone is never enough. A fire exit is quiet too.

A branch may want to split when its work falls into groups that different hands take up, or when a kind of work keeps coming that no accountability names. The read looks at who did what only to find the groups, then keeps the groups and lets the people go. What it says is the shape of the hands, never a hand: "nearly all of it done by the same person, a role in all but name", or "shared among three people, with nothing saying whose it is".

Two branches may want to merge when work keeps passing between them in both directions. Never because the same people hold both. Two hats on one head is no reason to fuse two jobs.

A branch that changed shape needs time before it's judged again. When a job has moved out into a branch of its own, or two have been joined, Wise-Eye offers nothing that would move it again for eight weeks. An offer someone let go isn't made again for four.

Every number here is a guess. We set them as placeholders and said so in the code. They'll change once real organisations live with them.

Where tensions land

The same traces show tensions, and each is read as the map is wrong here, never someone is doing too much or too little. There are five kinds the work shows by itself: work bouncing between two branches, one branch's work being done from another's seat, a branch keeping being sent work it doesn't hold, work arriving that no accountability answers for, and, since 1 November, overlap: the same people from outside coming to two branches, or promises to them split across two, so that nobody holds the whole of what they asked for.

Each lands on the lowest forum that covers every branch involved. A tension about one branch goes to the forum around that branch's parent, where its shape can change. A tension between two goes to the nearest forum they share. It arrives as a question, not a finding: where should the line between these two sit? Beside these sit the tensions people raised, the ones behind proposals that lapsed, were withdrawn, or whose review said they didn't work.

If a tension sits unanswered for fourteen days, it rises to the next forum up, and on up a forum a fortnight. At the root it says plainly that the whole organisation hasn't answered it yet. Someone seated where it sits can answer it in words, or raise a proposal from it.

The lean, and the four ways it reads

The read also asks where the organisation is heading. Most work is obligated: done in its seat, under its accountabilities, because it's the job. It shows the current shape and says nothing about desire. So the lean is read only from discretionary work, effort nobody had to spend: picking up what's outside your seat, taking in work with no home, acting on a direction the Field named, raising a proposal, calling a sensing, making a bet. Work Wise-Eye does never counts. Neither does anything on a branch's own board or in the Clearing.

Discretionary acts and what people said in the Field are placed in one shared set of themes. The voiced side fades, halving in thirty days. The revealed side is a movement, not a pile: this four weeks against the four before. It needs eight weeks of history and at least four discretionary acts in each stretch before it says anything, and until then the read says the work can't tell yet.

Then the two are set side by side, never blended. A strong lean is named and the chosen work is going there. Drift is named and the work hasn't moved. The unspoken is the work moving towards something, or thinning away from it, that nobody has named. And tested, not yet said is a theme where the organisation decided a bet, though no one has said it's where we're going. Each opens onto its evidence.

Drift and the unspoken do what the developmental psychologist Robert Kegan described as turning subject into object: something people were inside of, unseen, becomes something they can look at together.⁴ Neither is anyone's fault.

Otto Scharmer describes much of his work as helping a system to sense and see itself, so that people can act from the whole rather than from their own corner of it.⁵ He also names a blind spot: we study what leaders do and how they do it, and rarely the inner place they act from. The read can't see that place, and doesn't try. What it can do is hold the work up where everyone can look at it together.

The flame, and how it learns

Some of what an organisation needs to know isn't about shape. So each branch has a flame, read over itself and everything beneath it; the root's is the organisation's. It's made of four readings. Landing: promises kept, changes whose reviews said they worked, bets that held. On purpose: how the work sits with the organisation's own measures. Flowing: how far appreciation reaches across the organisation rather than round in corners. Warm: how full the organisation's gatherings have been.

The first two are about the work. The last two are about people, collectively, so they're read for a part of the tree only where at least five different people hold seats there. Below that, a branch's people count in the branch above, where no one can be told apart.

The flame is said in words, not shown as a number: burns bright, burns steady, flickers, burns low. While most of its readings are still on starting estimates, it says it's too soon to read, and on a branch that's gone quiet, that it's gone quiet. It flickers only when a reading it has actually read is straining. Branches are shown beside their siblings in the tree's own order, never ranked, and the report says at its foot that there are no goals and no nudges.

Beside it, for the whole organisation only, is a learning pulse: how long a tension takes to become a change, whether reviews are held and conclude, whether playbooks are revised as the work moves, whether directions the Field named turn into work. A signal that stays low is offered as something to look at.

The firewall

None of this decides anything. Nothing that makes decisions, moves trust, sets pay or seats a person draws on the read; the core that does those things can't reach it. The one place a reading travels is a budget proposal: when a forum is asked for money for a branch, the proposal carries what the read says of that branch, whether its work wants it to grow and its load of work, as it stood when the proposal was raised, so everyone deciding reads the same thing. It's evidence. The forum still decides.

How we decided

Much of what the read does today rests on instincts we wrote by hand: where work first goes, when a pass counts as a bounce, when a branch looks ready to divide. Chapter 21 describes where that's heading, a sense of itself the organisation learns from what organisations actually do, with the same firewall.

We chose to read flows rather than people, and to build the guard into the shape of the data rather than into a policy. A policy can be relaxed one exception at a time. A ledger with no place for a name can't.

We chose to weigh work by how it crossed the membrane because busyness looks exactly like need until you ask where the work came from.

We dropped an early rule that read two branches as wanting to merge when the same people held both. It confused the person with the role.

And we took a number off the branch page. The growth read, built to show us why structure did or didn't grow, gave each branch a line such as "none of six pieces in four weeks". People met it on their first day and read it as a quota to fill. It was Goodhart's law in miniature, inside our own design. The rule behind it still runs and still raises its offer when it's met; the figure now lives only where we study the system.

The flame was the harder call. We do score the organisation, even though people run it, because an organisation that can't see how it's doing can't tend itself. We hold that line by never scoring a person and by refusing every target.

What's still uncertain: every threshold, the length of the window, how fast the voiced lean should fade, and whether five is the right number of people before a reading stops being about someone. We also don't yet know whether people will feel a reading rising towards them as help or as pressure. We'll know when organisations live with it.

What you might take from this

An organisation can see itself without watching anyone. Read what the work does at each part, weigh it by whether the world asked, set what people say beside what they do without blending the two, and let no reading decide, pay or seat.

The work shows where the map is wrong, and it never needs to say whose fault that is.

Notes and further reading

On how measurement turns into its own goal: Muller, J. Z. (2018), The Tyranny of Metrics, Princeton University Press. On what happens when watching becomes the business: Zuboff, S. (2019), The Age of Surveillance Capitalism, Profile Books.